fix(detectors): update Dandelion token regex to match hex-only values

[diegmonti]

Description:

Change the regex pattern from [a-z0-9]{32} to [a-f0-9]{32} to only detect valid hexadecimal tokens, reducing false positives.

Checklist:

• Tests passing (make test-community)?
• Lint passing (make lint this requires golangci-lint)?

[CLAassistant]

All committers have signed the CLA.

[shahzadhaider1]

Hi @diegmonti,
Thank you for taking the time to open this PR. We really appreciate your contribution to the project.

To help reviewers confidently assess this change, could you please share any supporting documentation, specification, or authoritative reference confirming that Dandelion tokens are strictly hexadecimal ([a-f0-9]{32})?

Providing such context would make it easier for reviewers to verify the correctness of the change and ensure we’re not unintentionally missing valid tokens. Thanks again for your effort and support!

[vad]

Hi @shahzadhaider1, Dandelion developer here. You can check my organizations here on github, my role at https://www.spaziodati.eu/team and the Dandelion ownership in the footer of https://dandelion.eu/ .

We don’t strictly enforce tokens to be hexadecimal, but it’s also true that there are still some old tokens around that aren’t even 32 characters long, as detected by the current regex. All recent tokens follow exactly the format proposed by @diegmonti , so I’m fine with it.

Also, since the current detector implementation makes API calls to check whether a token is valid, we’ve noticed some request spikes (maybe a CI gone wild?) that we’d like to try to reduce.

[shahzadhaider1]

Hi @shahzadhaider1, Dandelion developer here. You can check my organizations here on github, my role at https://www.spaziodati.eu/team and the Dandelion ownership in the footer of https://dandelion.eu/ .

We don’t strictly enforce tokens to be hexadecimal, but it’s also true that there are still some old tokens around that aren’t even 32 characters long, as detected by the current regex. All recent tokens follow exactly the format proposed by @diegmonti , so I’m fine with it.

Also, since the current detector implementation makes API calls to check whether a token is valid, we’ve noticed some request spikes (maybe a CI gone wild?) that we’d like to try to reduce.

Grazie Davide, for the explanation.